An ultimate amp protection circuit ?

[Esperado]

here is the original schematic from the author and modified schematic using moc3041.

The R31 -> R37 value of 75.7K is not the good one. It has to be one of the values provided in the list, according temperature target.
WHat advantage did-you see, using moc3041 ? The original static relays wait for a 0 to allow switch of coils with no parasitic effects..

 

[coolet]

nothing... except that it has zero-cross detection and widely available worldwide.. and the diyers can easily select what triac to use.

 

[coolet]

ok i hope this is it

 

[Esperado]

What do you think for the offset of temperature alarm ? 10° ?

 

[coolet]

it was not the value of r37... it was a copy and paste error from r28 forgot to delete the value

already edit the schematic

 

[andrewlebon]

coolet you really working hard ON THE PROJECT

 

[coolet]

yes andrew.. i like the approach of esperado... comparing the input and output if i am not mistaken

 

[andrewlebon]

H i Coolet
greetings yes its seems like the ultimate protection not worried about the component count in my place some components are hard to find i am trying hard to get all
the components have to make this protection thanks to ESPERADO and you for your inputs
warm regards
andrew lebon

 

[coolet]

great i have a PVI5080 in place of OP3 and OP4 and NSL-32SR3, but NSL-32SR3 has a 60ohms when fully enabled, can it be used in place of OP1 and OP2?

 

[Esperado]

its seems like the ultimate protection not worried about the component count

In fact, the components count are not so much for each feature: not only DC and soft start as some pretend. And expecteded to be cheap.
Coolet, thanks again for your help, it encourage to drive this project to the end.

Here my proposal for hysteresis of the temperature. AT the start, as the temp is low, the CTN has a great value > than the Rtemp resistance: the out of the comparator is +12V.
When the CTN reach less than the value of Rtemp, the opamp will fire to -12V. Via the diode D, the + input, previously at 0V will go at ~ -2V. The OPA will stay in this state until the temp goes -10° under the temperature where it fired. Tuning the value of the 5K will change the ofset.

Thanks, whizgeek, for your suggestion.

 

[Esperado]

Coolet, and others, may-be you can have some benefit, reading this ?
ESP - MOSFET Solid State Relays
If i remember well, there is too a topic about mosfets relays for speakers in this forum. Will try to re-find it.
As i use only traditional relays, i have not much experience of them and, as i'm far from electronic business since a decade, it is difficult for me to find the good parts, both easily available and best for the use.
This is supposed to be a collaborative project. Will try to help the best i can, people with expertise in one or an other aspect are welcome, like the nice work of Coolet. We really need a good board designer. I have send a PM at Alex mm, but he seems away from the forum since some days.

 

[Esperado]

http://www.diyaudio.com/forums/solid-state/155939-mosfet-relays.html
http://www.diyaudio.com/forums/solid-state/191449-output-relays-23.html
etc...
I believe, when this part will be finished and good parts chosen, the electronic design will be near completed.

 

[Esperado]

Here the various solutions to chose to fire the mosfets.
http://www.geofex.com/Article_Folders/ampprot/dcprot2.gif
http://www.diyaudio.com/forums/attachments/solid-state/149566d1260281749-mosfet-relays-chocolv2.gif
http://www.diyaudio.com/forums/atta...56990d1324972075-output-relays-dcprot_mos.png


Photovoltaic Opto couplers for the mosfets: APV1121S, PV1050N, ASSR-V621-302E, PVI5013, TPLP 191B, TPLP591
Paralleling double ones for one channel will give faster switching time.


Power mosfet low rds non:
the best one for rails <+-50V
IPB025N10N3 G - MOSFET CANAL N 100V 180A TO263-7
for the same VDS:
IRF3710ZG 100V 0.018ohm 59A
IRFS59N10D 100V 0.025ohm 59A

for more Vrails:
IRF2907 IRFP2907 IRFP90N20D 200V 0.023ohm 94A - $4.55
IRFB4227PBF 200V 0.024ohm 65A
IRFB1227 200V 0.0197ohm 130A
IRFP90N20D 200V 0.023ohm 94A

May-be we can // two pairs of devices to reduce RDS on.

 

[jan.didden]

I build one last year based on an idea by Chocoholic, see attached.
I had a need for a fast ss relay for some test equipment so it uses a dedicated MOSFET driver.
IIRC it switches in just a few 100 nS.
The whole thing runs off 5V and is fully floating, can be used as a speaker relay, but also as power supply relay, up to several 100V (depending on specific MOSFET used).
The Rdson is just 10mohms or so and is quite linear with current so it does not introduce any distortion.

jan

 

[Esperado]

Thanks to join us, Jan.
I had read your schematic before. As the UAPC circuit is yet pretty complicated, i'm searching for a simpler solution to fire the mosfets. Any experience with simple photovoltaic opto couplers (listed above) ?

 

[ChocoHolic]

Direct drive from simple photovoltaic couplers cannot deliver much current and this leads to slow gate drive. Only for low power applications this solution will ensure switching inside the SOA of the MosFets.
Despite this there seem to be people who ignore SOA and did not report defects.

Jans approach is already close to the simplest solution, which provides proper gate drive. I think it could be simplified by reducing the pos/neg supply to a single positive supply only, but this will save just one e-cap.

Hm, with photovoltaic couplers, there are two improvements possible.
a) Speeding up the turn OFF needs just an additional PNP, while the turn ON could be kept slow through a diode. So additional one diode and one PNP will already give a reasonable turn OFF.
b) Ensuring proper gate drive for turn ON would need to charge a floating supply cap first and after the cap has be charged then turning on the MosFet with a NPN. Unfortunately the complexity including a uP-comparator would by far exceed the complexity of Jans approach.

==> With respect to the fact that turning OFF in case the amp is defect and the speaker has to be protected is the most important and demanding situation - chosing a photovoltaic coupler and enhance it according a) could be reasonable trade off.

 

[Esperado]

Thanks a lot, ChocoHolic, for the perfect summary, just what i expected
Did-you have some values of power off switching times (power on has no importance) using two side of double photovolatic ?
In our present situation, No voltage is supposed to be present at power on/off during normal situation.

 

[andrewlebon]

Hi Alex mm
greetings you are at the right place after a long time
warm regards
andrew lebon

 

[Bigun]

Direct drive from simple photovoltaic couplers cannot deliver much current and this leads to slow gate drive. Only for low power applications this solution will ensure switching inside the SOA of the MosFets.
Despite this there seem to be people who ignore SOA and did not report defects.

Jans approach is already close to the simplest solution, which provides proper gate drive. I think it could be simplified by reducing the pos/neg supply to a single positive supply only, but this will save just one e-cap.

Hm, with photovoltaic couplers, there are two improvements possible.
a) Speeding up the turn OFF needs just an additional PNP, while the turn ON could be kept slow through a diode. So additional one diode and one PNP will already give a reasonable turn OFF.
b) Ensuring proper gate drive for turn ON would need to charge a floating supply cap first and after the cap has be charged then turning on the MosFet with a NPN. Unfortunately the complexity including a uP-comparator would by far exceed the complexity of Jans approach.

==> With respect to the fact that turning OFF in case the amp is defect and the speaker has to be protected is the most important and demanding situation - chosing a photovoltaic coupler and enhance it according a) could be reasonable trade off.

Judging by both the manufacturers data sheet for the photovoltaic coupler that I used and the fact that my own implementation works just fine I believe these photovoltaic couplers can provide plenty of drive for this application. Turn-on time is a non-issue of course because a small delay in turning on is of no consequence to anybody. The turn-off time is of consequence and it is normal to add a resistor in parallel with the output of the photovoltaic coupler to provide a path for the FET gate capacitors to discharge. This is the recommendation of the photovoltaic manufacturer. The resistor has to be properly sized, too big and the turn-off is slower than it could be otherwise, too small and there won't be enough voltage to turn-on. It wasn't too hard to calculate the right value. In my view this approach is sufficient for a dc-protection device because the time for the FETs to turn-off is still short compared with the time it takes the dc-detection circuit to integrate the signal and 'determine' that a dc-error is present. So the time from dc-error condition to turn-off is dominated by the dc-error detection circuit and not the solid state relay.

I didn't want to introduce any more components than necessary as the fewest number of parts often means the most reliable.

I tested mine with a real speaker (not my most expensive one mind) and it works fine. There is no impact on sound quality. And of most importance - I trust it.

The one thing I forgot to include in my design was an ac-detect to shut off the speakers when the power to the amplifier is turned off - this would avoid any turn-off thumps. The problem with my amplifier is that the output does drift quite a bit when the amplifier is turned off, but it does so slowly so you'd never hear it. But with my dc-error detection circuit in place the output hits the dc-error threshold and my protection circuit cuts off the speaker and this makes it thump. So, ironically the dc-protect circuit is causing a turn-off thump which the amplifier doesn't generate on it's own. My advice - install additional circuity to cut off the speakers when you power down the amplifier.

 

[Esperado]

Hi, bigun, nice to see-you here,
Did-you have an idea of your off switching time ?
I was very surprised to see no mentions of the off resistance of the various Photovoltaic switches in data-sheets, except two, including one at 30nS with 1nF of mosfet capacitance. Data *shiits* are so poor nowadays, that is is quite impossible to work with them !!!